CN115795255B - Method, device, medium and terminal for detecting time sequence change of wetland - Google Patents
Method, device, medium and terminal for detecting time sequence change of wetland Download PDFInfo
- Publication number
- CN115795255B CN115795255B CN202211152897.0A CN202211152897A CN115795255B CN 115795255 B CN115795255 B CN 115795255B CN 202211152897 A CN202211152897 A CN 202211152897A CN 115795255 B CN115795255 B CN 115795255B
- Authority
- CN
- China
- Prior art keywords
- change
- wetland
- data
- band data
- denoised
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 230000008859 change Effects 0.000 title claims abstract description 330
- 238000000034 method Methods 0.000 title claims abstract description 47
- 238000004364 calculation method Methods 0.000 claims abstract description 21
- 238000012216 screening Methods 0.000 claims abstract description 17
- 238000001514 detection method Methods 0.000 claims description 29
- 238000006243 chemical reaction Methods 0.000 claims description 16
- 238000012545 processing Methods 0.000 claims description 15
- 238000004891 communication Methods 0.000 claims description 10
- 238000001914 filtration Methods 0.000 claims description 4
- 238000000605 extraction Methods 0.000 claims description 3
- 230000008569 process Effects 0.000 description 12
- 238000002310 reflectometry Methods 0.000 description 8
- 230000006870 function Effects 0.000 description 6
- 230000004913 activation Effects 0.000 description 5
- 230000015556 catabolic process Effects 0.000 description 5
- 238000006731 degradation reaction Methods 0.000 description 5
- 238000011160 research Methods 0.000 description 4
- 239000011159 matrix material Substances 0.000 description 3
- 238000012546 transfer Methods 0.000 description 3
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 3
- 230000007547 defect Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- IGRCWJPBLWGNPX-UHFFFAOYSA-N 3-(2-chlorophenyl)-n-(4-chlorophenyl)-n,5-dimethyl-1,2-oxazole-4-carboxamide Chemical compound C=1C=C(Cl)C=CC=1N(C)C(=O)C1=C(C)ON=C1C1=CC=CC=C1Cl IGRCWJPBLWGNPX-UHFFFAOYSA-N 0.000 description 1
- 208000005156 Dehydration Diseases 0.000 description 1
- 241001465754 Metazoa Species 0.000 description 1
- 230000001133 acceleration Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000006378 damage Effects 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 230000035772 mutation Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000009304 pastoral farming Methods 0.000 description 1
- 238000012958 reprocessing Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 230000003595 spectral effect Effects 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
Abstract
The invention discloses a method, a device, a medium and a terminal for detecting time sequence change of a wetland, wherein the method comprises the steps of screening time sequence data to obtain wetland wave band data, and carrying out denoising treatment on the wetland wave band data to obtain denoised wetland wave band data; inputting the denoised wetland band data into an LSTM model for fitting calculation to obtain fitting wetland band data; performing difference comparison on the fitted wetland band data and the denoised wetland band data to obtain change disturbance information, calculating the change probability of wetland change and the change angle of the change disturbance information, and performing change judgment according to the change probability and the change angle to obtain a judgment result; the method and the device select to screen the change disturbance information according to the judgment result to obtain the wetland time sequence change information.
Description
Technical Field
The present invention relates to the field of wetland detection, and in particular, to a method, an apparatus, a medium, and a terminal for detecting time series changes of a wetland.
Background
Currently, the detection and research of the change of the wetland are usually performed by adopting a transfer matrix mode, and when the change result of the wetland is analyzed by adopting the transfer matrix mode, various analysis results are usually obtained, such as: 1. the method has the advantages that the wetland change result of the local area shows that the Chinese wetland overall shows a remarkable degradation trend, and degradation forms show diversification, so that the area degradation exists, and the type of evolution degradation exists; 2. the change research result of the whole wetland shows that the wetland is still in an overall reduced state despite the increase of the wetland area, and the main change situation of the wetland has obvious stage; 3. the change characteristics of the Chinese wetland have obvious regional differences in space, and can be divided into three modules, the wetland in the western region shows an increasing trend, the conditions of increasing and decreasing in the middle region coexist, and the wetland in the eastern region is a degradation trend. For various analysis results obtained in the above-mentioned wetland study, on one hand, it is considered that the change of the wetland may be actually the same, and on the other hand, the change result is obtained by calculating the difference value of the data of the first two years by adopting a transfer matrix mode, and because the result error fluctuation of the wetland data of the two years is larger and the accidental error of the change detection result is higher, the obtained wetland change detection result will lose a large amount of effective information in the middle of the two years, thereby reducing the accuracy and the robustness of the analysis result.
Therefore, we propose a time series change detection method of the wetland to improve the precision of the wetland change result, thereby improving the accuracy and the robustness of the wetland analysis result.
Disclosure of Invention
In view of the above-mentioned shortcomings of the prior art, an object of the present application is to provide a method, an apparatus, a medium and a terminal for detecting time-series changes of a wetland, which aim to solve the problem of how to improve the accuracy of the result of the wetland change.
In order to solve the above technical problems, a first aspect of embodiments of the present application provides a method for detecting a time series change of a wetland, where the method includes:
filtering the time sequence data to obtain wetland band data, and denoising the wetland band data to obtain denoised wetland band data;
inputting the denoised wetland band data into an LSTM model for fitting calculation to obtain fitting wetland band data;
performing difference comparison on the fitted wetland band data and the denoised wetland band data to obtain change disturbance information, calculating the change probability of wetland change and the change angle of the change disturbance information, and performing change judgment according to the change probability and the change angle to obtain a judgment result;
and selecting and screening the change disturbance information according to the judging result to obtain wetland time sequence change information.
As a further improved technical solution, the filtering the time-series data to obtain wetland band data, and denoising the wetland band data to obtain denoised wetland band data includes:
performing band data screening on Landsat time sequence data to obtain a plurality of wetland band sub-data, and combining the plurality of wetland band sub-data to obtain the wetland band data, wherein the wetland band sub-data comprises a Red band, an NIR band, a SWIR1 band, an NDVI index, an MNDWI index, a TCB (ternary content addressable System) leaf cap conversion index, a TCG leaf cap conversion index, a TCW leaf cap conversion index and a TCA leaf cap conversion index;
performing preliminary denoising treatment on the wetland band sub-data by adopting an FMASK algorithm to obtain preliminary treatment data;
fitting the preliminary processing data by adopting a RIRLS formula to obtain a fitting value;
and denoising again based on the fitting value and the dynamic threshold value between the preliminary processing data to obtain denoised wetland band sub-data, and combining a plurality of denoised wetland band sub-data to obtain denoised wetland band data.
As a further improved technical solution, the rills formula is:
wherein, in the formulaFor fitting values, x is the date, i is the ith wetland band sub-data, T is the number of days per year, N is the number of years of the wetland band sub-data, a 0,i A is the total variation coefficient value of the ith wetland wave band sub-data 1,i And b 1,i All are annual change coefficients, a 2,i And b 2,i All are annual change coefficients.
As a further improved technical scheme, the step of inputting the denoised wetland band data into an LSTM model for fitting calculation, where obtaining the fitting wetland band data includes:
and inputting the denoised wetland band data into an LSTM model for fitting calculation to obtain the fitting wetland band data, wherein the LSTM model comprises an input gate, a forgetting gate, an output gate and a long and short memory gate.
As a further improved technical scheme, the difference comparison is performed between the fitted wetland band data and the denoised wetland band data to obtain variation disturbance information, variation probability of wetland variation and variation angle of the variation disturbance information are calculated, variation judgment is performed according to the variation probability and the variation angle, and the obtained judgment result comprises:
performing difference comparison on the fitted wetland band data and the denoised wetland band data to obtain the variation disturbance information of the denoised wetland band data;
calculating the probability of wetland change based on the difference value between the fitted wetland band data and the denoised wetland band data to obtain the change probability value, and comparing the change probability value with a dynamic threshold value;
calculating the change angle of the change disturbance information to obtain a change angle value, and comparing the change angle value with an angle threshold value;
and carrying out change judgment according to the change probability and the change angle to obtain a changed judgment result and an unchanged judgment result.
As a further improved technical solution, the performing the change determination according to the change probability and the change angle, and obtaining the determination result includes:
if the change probability value is larger than the dynamic threshold value and the change angle value is larger than the angle threshold value, the change is judged, and a change judgment result is obtained;
and if the change probability value is smaller than the dynamic threshold value and/or the change angle value is smaller than the angle threshold value, judging that the change is unchanged, and obtaining an unchanged judging result.
As a further improved technical solution, the selecting, according to the determination result, the change disturbance information to screen, to obtain wetland time sequence change information includes:
if the unchanged judging result is obtained, selecting not to screen the changed disturbance information, and taking the changed disturbance information as the wetland time sequence change information;
and if the judgment result of the change is obtained, selecting to screen the change disturbance information, and removing vegetation green degree increasing information caused by vegetation turning green and artificial factors in the change disturbance information to obtain the wetland time sequence change information.
A second aspect of the embodiments of the present application provides a time-series change detection device for a wetland, including:
the data processing module is used for screening the time sequence data to obtain wetland band data, and carrying out denoising treatment on the wetland band data to obtain denoised wetland band data;
the data fitting module is used for inputting the denoised wetland band data into an LSTM model for fitting calculation to obtain fitting wetland band data;
the change detection module is used for carrying out difference comparison on the fitted wetland wave band data and the denoised wetland wave band data to obtain change disturbance information, calculating the change probability of wetland change and the change angle of the change disturbance information, and carrying out change judgment according to the change probability and the change angle to obtain a judgment result;
and the disturbance information extraction module is used for selecting the change disturbance information according to the judging result to obtain wetland time sequence change information.
A third aspect of the embodiments of the present application provides a computer-readable storage medium storing one or more programs executable by one or more processors to implement steps in a method for detecting a time-series change of a wetland as described in any one of the above.
A fourth aspect of the present embodiment provides a terminal device, including: a processor, a memory, and a communication bus; the memory has stored thereon a computer readable program executable by the processor;
the communication bus realizes connection communication between the processor and the memory;
the processor, when executing the computer readable program, implements the steps in the method for detecting time series change of wetland as described in any one of the above.
The beneficial effects are that: compared with the prior art, the time sequence change detection method of the wetland comprises the steps of screening time sequence data to obtain wetland band data, and carrying out denoising treatment on the wetland band data to obtain denoised wetland band data; inputting the denoised wetland band data into an LSTM model for fitting calculation to obtain fitting wetland band data; performing difference comparison on the fitted wetland band data and the denoised wetland band data to obtain change disturbance information, calculating the change probability of wetland change and the change angle of the change disturbance information, and performing change judgment according to the change probability and the change angle to obtain a judgment result; according to the method, the defect of lack of wetland characteristics and fitting parameter dependence in a current change detection model can be overcome, accurate wetland time sequence change information can be obtained, and the accuracy and the robustness of a wetland analysis result can be improved.
Drawings
Fig. 1 is a flowchart of a method for detecting time-series change in a wet land according to the present invention.
Fig. 2 is a schematic structural diagram of a terminal device provided by the present invention.
Fig. 3 is a block diagram of the structure of the device provided by the invention.
Fig. 4 is a flow chart of a method provided by the invention.
The achievement of the objects, functional features and advantages of the present invention will be further described with reference to the accompanying drawings, in conjunction with the embodiments.
Detailed Description
In order to facilitate an understanding of the present application, a more complete description of the present application will now be provided with reference to the relevant figures. Preferred embodiments of the present application are shown in the accompanying drawings. This application may, however, be embodied in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein in the description of the application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application.
The inventors have found that the following problems exist in the prior art:
(1) Along with the acceleration of the global change speed, the response speed of the ecological system to the global change is also accelerated, new requirements are put forward for the time resolution of the land cover data, the trend of the time sequence of the land cover data is expressed as the change process from 10 years scale to 5 years scale to year scale to season scale to month scale, the time resolution of the land cover data is improved, the effect of reducing accidental errors of change detection is obvious, the wetland ecological system is sensitive to the quick change environment, the time-space change characteristics of the wetland ecological system are detected by using the long-time sequence data of the wetland, and the uncertainty result caused by the accidental errors is hopefully reduced. At present, there are researches on time sequence changes in aspects of abandoned farmland, disturbance of forest lands, expansion of construction lands, lakes and the like. The wetland has high difficulty in detecting changes due to inherent complex characteristics and the like, and the wetland is lack of researches on targeted long-time sequence changes at present.
In order to solve the above problems, various non-limiting embodiments of the present application are described in detail below with reference to the accompanying drawings.
As shown in fig. 1, the method for detecting the time sequence change of the wetland provided by the embodiment of the application includes the following steps:
s1, screening time sequence data to obtain wetland band data, and carrying out denoising treatment on the wetland band data to obtain denoised wetland band data;
specifically, because the wetland mainly comprises two elements of vegetation and water, the wetland change detection needs to be sensitive to the two elements at the same time, firstly, data of the two elements of vegetation and water in Landsat time sequence data are screened out to obtain wetland band data, then, the wetland band data are subjected to noise pretreatment and noise reprocessing in sequence, noise such as clouds and shadows in the wetland band data is removed, and finally, the denoised wetland band data are obtained.
The method comprises the steps of screening time sequence data to obtain wetland band data, denoising the wetland band data to obtain denoised wetland band data, and the steps of:
s101, performing band data screening on Landsat time sequence data to obtain a plurality of wetland band sub-data, and combining the plurality of wetland band sub-data to obtain the wetland band data, wherein the wetland band sub-data comprises a Red band, an NIR band, a SWIR1 band, an NDVI index, an MNDWI index, a TCB leaf cap conversion index, a TCG leaf cap conversion index, a TCW leaf cap conversion index and a TCA leaf cap conversion index;
s102, performing preliminary denoising treatment on the wetland wave band sub-data by adopting an FMASK algorithm to obtain preliminary treatment data;
s103, fitting the preliminary processing data by adopting a RIRLS formula to obtain a fitting value;
and S104, denoising again based on the fitting value and the dynamic threshold value between the preliminary processing data to obtain denoised wetland band sub-data, and combining a plurality of denoised wetland band sub-data to obtain denoised wetland band data.
The rills formula is:
wherein, in the formulaFor fitting values, x is the date, i is the ith wetland band sub-data, T is the number of days per year, N is the number of years of the wetland band sub-data, a 0,i A is the total variation coefficient value of the ith wetland wave band sub-data 1,i And b 1,i All are annual change coefficients, a 2,i And b 2,i All are annual change coefficients.
Specifically, band data screening is performed on Landsat time sequence data to obtain a plurality of pieces of wetland band sub-data, the wetland band sub-data in this embodiment includes Red band, NIR band, SWIR1 band, NDVI (normalized difference vegetation index) index, MNDWI (modified normalized difference water index) index, TCB (Tasseled Cap brightness) leaf-cap conversion index, TCG (Tasseled Cap greenness) leaf-cap conversion index, TCW (Tasseled Cap Wetness) leaf-cap conversion index and TCA (Tasseled Cap Angle) leaf-cap conversion index, the nine pieces of wetland band sub-data are combined to obtain wetland band data, because the change detection process is sensitive to noise data of an image, interference information related to external non-wetland change is required to be accurately removed, firstly, the nine pieces of wetland band sub-data are subjected to preliminary denoising processing by adopting an FMASK algorithm to obtain nine pieces of preliminary processing data, then fitting is performed on the nine pieces of preliminary processing data sequentially by adopting a robust iterative weighted least square (RIS) formula to obtain nine pieces of values, the nine pieces of preliminary processing data are respectively corresponding to the nine pieces of fitting values, denoising is performed again according to a dynamic threshold between the fitting values and the corresponding preliminary processing data, the following specific data is obtained, the specific data is subjected to denoising process is determined according to a specific data value domain of the following data, and the specific data is removed, and the specific data domain is determined after the following data is subjected to denoising process is determined:
in the above-mentioned method, the step of,x represents a date for a value range of specific data to be removed; ρ (i, x) is the preliminary processed data value, +.>For fitting values +.>And finally, combining the nine denoised wetland band sub-data to obtain denoised wetland band data.
S2, inputting the denoised wetland band data into an LSTM model for fitting calculation to obtain fitting wetland band data;
specifically, the result of the change detection is mainly judged by judging the difference between the actual wetland band data and the wetland band data after model fitting, so that the accuracy of model fitting is directly related to the result of the change detection, the LSTM model is a non-parametric model, the calculation amount is reduced compared with a harmonic regression method adopted in CCDC and COLD, the calculation amount is required to be reduced through the process of determining parameters through LASSO regression, the calculation efficiency of the model is effectively improved, and the calculation efficiency is also higher than the Kalman filtering of a state space model, so that the denoised wetland band data is input into the LSTM model for fitting calculation, and the fitted wetland band data is obtained.
The step of inputting the denoised wetland band data into an LSTM model for fitting calculation, and the step of obtaining fitting wetland band data comprises the following steps:
and inputting the denoised wetland band data into an LSTM model for fitting calculation to obtain the fitting wetland band data, wherein the LSTM model comprises an input gate, a forgetting gate, an output gate and a long and short memory gate.
Specifically, the LSTM model formula is as follows:
in the LSTM model formula, f is the activation function of a gate, g is the activation function of the Cell input, h is the activation function of the Cell output, sigma is the Sigmoid activation function, and tanh is the activation function, W i Is the weight of the i gate.
S3, performing difference comparison on the fitted wetland band data and the denoised wetland band data to obtain change disturbance information, calculating the change probability of wetland change and the change angle of the change disturbance information, and performing change judgment according to the change probability and the change angle to obtain a judgment result;
specifically, after the fitted wetland band data is obtained, the fitted wetland band data and the denoised wetland band data are required to be subjected to difference comparison to obtain change disturbance information about the wetland, then whether the wetland is changed or not is judged and judged, if the wetland is not changed, the change disturbance information is directly output, and if the wetland is changed, vegetation green degree increase information caused by vegetation green change or other human factors in the change disturbance information is required to be further removed, so that the accuracy of the change disturbance information is improved.
The method comprises the steps of performing difference comparison on the fitted wetland band data and the denoised wetland band data to obtain change disturbance information, calculating the change probability of wetland change and the change angle of the change disturbance information, and performing change judgment according to the change probability and the change angle to obtain a judgment result, wherein the judgment result comprises the following steps of:
s301, performing difference comparison on the fitted wetland band data and the denoised wetland band data to obtain the change disturbance information of the denoised wetland band data;
s302, calculating the probability of wetland change based on the difference value between the fitted wetland band data and the denoised wetland band data to obtain the change probability value, and comparing the change probability value with a dynamic threshold value;
s303, calculating the change angle of the change disturbance information to obtain a change angle value, and comparing the change angle value with an angle threshold;
s304, carrying out change judgment according to the change probability and the change angle to obtain a changed judgment result and an unchanged judgment result.
Wherein, the change judgment is carried out according to the change probability and the change angle, and the judgment result is obtained by the following steps:
s3041, if the change probability value is larger than the dynamic threshold value and the change angle value is larger than the angle threshold value, determining that the change is performed, and obtaining a change determination result;
and S3042, judging that the dynamic threshold value is unchanged if the change probability value is smaller than the dynamic threshold value and/or the change angle value is smaller than the angle threshold value, and obtaining an unchanged judgment result.
Specifically, firstly, comparing and finding the difference between the fitted wetland band data and the denoised wetland band data to obtain change disturbance information of the denoised wetland band data, wherein the change disturbance information is the difference information between the fitted wetland band data and the denoised wetland band data, then calculating the probability of wetland change based on the difference between the fitted wetland band data and the denoised wetland band data to obtain a change probability value, and comparing the change probability value with a dynamic threshold value, wherein the specific process is as follows:
wherein i is the ith wetland band, k is the wetland band data for performing change detection, ρ i For the denoised wetland band data,to fit wetland band data, +.>Representing a dynamic threshold;
then calculating the change angle of the change disturbance information to obtain a change angle value, comparing the change angle value with an angle threshold value, and keeping the change disturbance information relatively consistent in the change intensity and the change angle, wherein although certain noise data and non-wetland change information possibly exist in certain time sequences, the possibility that all the error information keeps consistent change directions is extremely low, so that the change angle needs to be added as a standard for judging the change, and the specific process is as follows:
in the above formula, i represents the i-th continuous change observation; k represents the number of continuous observations of the change that are confirmed to be true;
finally, carrying out change judgment according to the change probability and the change angle to obtain a changed judgment result and an unchanged judgment result, specifically, judging that the change is caused if the change probability value is larger than a dynamic threshold value and the change angle value is larger than an angle threshold value, and obtaining a changed judgment result; if the variation probability value is smaller than the dynamic threshold value and/or the variation angle value is smaller than the angle threshold value, the judgment is unchanged, and unchanged judgment results are obtained, wherein the angle threshold value is preferably 45 degrees.
And S4, selecting and screening the change disturbance information according to the judging result to obtain wetland time sequence change information.
Because typical wetlands are disturbed, lower NIR band reflectivity, higher RED band reflectivity and higher SWIR1 reflectivity values result, and these spectral changes may occur asynchronously, such as when the wetlands are disturbed by grazing activity or wild animals, the SWIR1 band reflectivity tends to increase, as water stress increases, the RED band reflectivity will also increase, and finally the NIR band reflectivity decreases due to the destruction of the wetland vegetation.
In addition, the mutation extracted from the wetland band data in the change detection process may not be related to wetland disturbance, some change points may be vegetation greenness increase information caused by vegetation itself change or other artificial factors, in order to eliminate the series of error information, the change disturbance information needs to be further screened on the basis of change detection, vegetation greenness increase information caused by vegetation change and artificial factors in the change disturbance information is removed, and finally wetland time sequence change information is obtained.
The step of selecting the change disturbance information according to the judging result to obtain wetland time sequence change information comprises the following steps:
s401, if the unchanged judging result is obtained, selecting not to screen the change disturbance information, and taking the change disturbance information as the wetland time sequence change information;
and S402, if the judgment result of the change is obtained, selecting to screen the change disturbance information, and removing vegetation green degree increasing information caused by vegetation turning green and artificial factors from the change disturbance information to obtain the wetland time sequence change information.
Specifically, whether to screen is firstly determined according to the obtained determination result, if the obtained unchanged determination result is not changed, the change disturbance information is selected not to screen and is directly output as wetland time sequence change information, if the changed determination result is obtained, the change disturbance information is selected to screen, vegetation greenness increasing information caused by vegetation greenness and artificial factors in the change disturbance information is removed, and finally the wetland time sequence change information is obtained.
If a variation judgment result is obtained, firstly, judging whether the variation belongs to wetland disturbance according to a green degree direction by comparing a variation vector obtained by model fitting and an actual spectrum value, and secondly, further judging a part with undefined disturbance information according to a slope trend direction, thereby determining final wetland disturbance information, wherein when the variation judgment result is implemented, the process of removing variation characteristics caused by vegetation turning green from the variation disturbance information comprises the following steps: according to the vegetation greening, high NIR reflectivity value and low Red and SWIR1 reflectivity values are generated, so that non-wetland disturbance change characteristics caused by vegetation greening can be screened out through a greenness direction formula, and the greenness direction formula is as follows:
ΔRed<-Threshold&ΔNIR>Threshold&ΔSWIR1<-Threshold
in addition, the characteristic of turning green is also provided with vegetation green degree increasing information caused by artificial wetland protection and the like, the forward disturbance information is extracted through the change of trend direction, and the forward disturbance information can be specifically screened through a slope trend direction formula, wherein the slope trend direction formula is as follows:
S Red,Aft <|S Red,Bef |&S NIR,Aft >|S NRIR,Bef |&S SWIR1,Aft <|S SWIR1,Bef |
and removing vegetation green degree increasing information caused by vegetation green and artificial factors from the change disturbance information after screening, finally obtaining wetland time sequence change information, and taking the wetland time sequence change information as a change detection result.
In summary, the invention optimizes the monitored characteristics and the model fitting process, replaces the common harmonic fitting and Least Absolute Shrinkage and Selection Operator (LASSO) method with large calculation amount by using the non-parametric LSTM network model, thereby achieving the purpose of improving the efficiency and the change detection precision.
Based on the above-mentioned method for detecting time-series changes of the wetland, the present embodiment provides a device for detecting time-series changes of the wetland, including:
the data processing module 1 is used for screening the time sequence data to obtain wetland band data, and carrying out denoising treatment on the wetland band data to obtain denoised wetland band data;
the data fitting module 2 is used for inputting the denoised wetland band data into an LSTM model for fitting calculation to obtain fitting wetland band data;
the change detection module 3 is used for performing difference comparison on the fitted wetland wave band data and the denoised wetland wave band data to obtain change disturbance information, calculating the change probability of wetland change and the change angle of the change disturbance information, and performing change judgment according to the change probability and the change angle to obtain a judgment result;
and the disturbance information extraction module 4 is used for selecting the change disturbance information according to the judging result to screen so as to obtain wetland time sequence change information.
In addition, it should be noted that the working process of the wetland-based time series change detection device provided in this embodiment is the same as that of the above-mentioned wetland time series change detection method, and specifically, reference may be made to the working process of the wetland time series change detection method, which is not described herein again.
Based on the above-described method for detecting time-series variation of a wetland, the present embodiment provides a computer-readable storage medium storing one or more programs executable by one or more processors to implement the steps in the method for detecting time-series variation of a wetland as described in the above-described embodiment.
As shown in fig. 2, based on the above-mentioned method for detecting time-series variation of wetland, the present application also provides a terminal device, which includes at least one processor (processor) 20; a display screen 21; and a memory (memory) 22, which may also include a communication interface (Communications Interface) 23 and a bus 24. Wherein the processor 20, the display 21, the memory 22 and the communication interface 23 may communicate with each other via a bus 24. The display screen 21 is configured to display a user guidance interface preset in the initial setting mode. The communication interface 23 may transmit information. The processor 20 may invoke logic instructions in the memory 22 to perform the methods of the embodiments described above.
Further, the logic instructions in the memory 22 described above may be implemented in the form of software functional units and stored in a computer readable storage medium when sold or used as a stand alone product.
The memory 22, as a computer readable storage medium, may be configured to store a software program, a computer executable program, such as program instructions or modules corresponding to the methods in the embodiments of the present disclosure. The processor 20 performs functional applications and data processing, i.e. implements the methods of the embodiments described above, by running software programs, instructions or modules stored in the memory 22.
The memory 22 may include a storage program area that may store an operating system, at least one application program required for functions, and a storage data area; the storage data area may store data created according to the use of the terminal device, etc. In addition, the memory 22 may include high-speed random access memory, and may also include nonvolatile memory. For example, a plurality of media capable of storing program codes such as a usb disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (Random Access Memory, RAM), a magnetic disk, or an optical disk, or a transitory storage medium may be used.
Compared with the prior art, the time sequence change detection method of the wetland comprises the steps of screening time sequence data to obtain wetland band data, and carrying out denoising treatment on the wetland band data to obtain denoised wetland band data; inputting the denoised wetland band data into an LSTM model for fitting calculation to obtain fitting wetland band data; performing difference comparison on the fitted wetland band data and the denoised wetland band data to obtain change disturbance information, calculating the change probability of wetland change and the change angle of the change disturbance information, and performing change judgment according to the change probability and the change angle to obtain a judgment result; according to the method, the defect of lack of wetland characteristics and fitting parameter dependence in a current change detection model can be overcome, accurate wetland time sequence change information can be obtained, and the accuracy and the robustness of a wetland analysis result can be improved.
It is to be understood that the invention is not limited in its application to the examples described above, but is capable of modification and variation in light of the above teachings by those skilled in the art, and that all such modifications and variations are intended to be included within the scope of the appended claims.
The above examples of the present invention are of course more detailed, but should not be construed as limiting the scope of the invention, and various other embodiments are possible, based on which those skilled in the art can obtain other embodiments without any inventive task, which fall within the scope of the invention as defined in the appended claims.
Claims (7)
1. A method for detecting time-series change of a wet land, comprising:
filtering the time sequence data to obtain wetland band data, and denoising the wetland band data to obtain denoised wetland band data;
inputting the denoised wetland band data into an LSTM model for fitting calculation to obtain fitting wetland band data;
performing difference comparison on the fitted wetland band data and the denoised wetland band data to obtain change disturbance information, calculating the change probability of wetland change and the change angle of the change disturbance information, and performing change judgment according to the change probability and the change angle to obtain a judgment result;
and performing difference comparison on the fitted wetland band data and the denoised wetland band data to obtain change disturbance information, calculating the change probability of wetland change and the change angle of the change disturbance information, and performing change judgment according to the change probability and the change angle to obtain a judgment result, wherein the judgment result comprises:
performing difference comparison on the fitted wetland band data and the denoised wetland band data to obtain the variation disturbance information of the denoised wetland band data;
calculating the probability of wetland change based on the difference value between the fitted wetland band data and the denoised wetland band data to obtain the change probability value, and comparing the change probability value with a dynamic threshold value;
calculating the change angle of the change disturbance information to obtain a change angle value, and comparing the change angle value with an angle threshold value;
carrying out change judgment according to the change probability and the change angle to obtain a changed judgment result and an unchanged judgment result;
and performing change judgment according to the change probability and the change angle to obtain a judgment result, wherein the step of obtaining the judgment result comprises the following steps of:
if the change probability value is larger than the dynamic threshold value and the change angle value is larger than the angle threshold value, the change is judged, and a change judgment result is obtained;
if the change probability value is smaller than the dynamic threshold value and/or the change angle value is smaller than the angle threshold value, judging that the change is unchanged, and obtaining an unchanged judging result;
the step of selecting the change disturbance information according to the judging result to obtain wetland time sequence change information comprises the following steps:
if the unchanged judging result is obtained, selecting not to screen the changed disturbance information, and taking the changed disturbance information as the wetland time sequence change information;
if the judging result of the change is obtained, judging the part of the change disturbance information which is not clear through the slope trend direction, determining final change disturbance information, removing vegetation green degree increasing information caused by vegetation green and artificial factors in the change disturbance information, obtaining wetland time sequence change information, and taking the wetland time sequence change information as a change detection result;
and selecting and screening the change disturbance information according to the judging result to obtain wetland time sequence change information.
2. The method for detecting time series variation of a wetland according to claim 1, wherein the step of screening the time series data to obtain wetland band data, the step of denoising the wetland band data to obtain denoised wetland band data comprises the steps of:
performing band data screening on Landsat time sequence data to obtain a plurality of wetland band sub-data, and combining the plurality of wetland band sub-data to obtain the wetland band data, wherein the wetland band sub-data comprises a Red band, an NIR band, a SWIR1 band, an NDVI index, an MNDWI index, a TCB (ternary content addressable System) leaf cap conversion index, a TCG leaf cap conversion index, a TCW leaf cap conversion index and a TCA leaf cap conversion index;
performing preliminary denoising treatment on the wetland band sub-data by adopting an FMASK algorithm to obtain preliminary treatment data;
fitting the preliminary processing data by adopting a RIRLS formula to obtain a fitting value;
and denoising again based on the fitting value and the dynamic threshold value between the preliminary processing data to obtain denoised wetland band sub-data, and combining a plurality of denoised wetland band sub-data to obtain denoised wetland band data.
3. The method for detecting time-series variation of a wetland according to claim 2, wherein the rills formula is:
wherein, in the formulaFor fitting values, x is the date, i is the ith wetland band sub-data, T is the number of days per year, N is the number of years of the wetland band sub-data, a 0,i A is the total variation coefficient value of the ith wetland wave band sub-data 1,i And b 1,i All are annual change coefficients, a 2,i And b 2,i All are annual change coefficients.
4. The method for detecting time series variation of wetland according to claim 3, wherein said inputting the denoised wetland band data into an LSTM model for fitting calculation, obtaining fitted wetland band data comprises:
and inputting the denoised wetland band data into an LSTM model for fitting calculation to obtain the fitting wetland band data, wherein the LSTM model comprises an input gate, a forgetting gate, an output gate and a long and short memory gate.
5. A time-series change detection device for a wet land, comprising:
the data processing module is used for screening the time sequence data to obtain wetland band data, and carrying out denoising treatment on the wetland band data to obtain denoised wetland band data;
the data fitting module is used for inputting the denoised wetland band data into an LSTM model for fitting calculation to obtain fitting wetland band data;
the change detection module is used for carrying out difference comparison on the fitted wetland wave band data and the denoised wetland wave band data to obtain change disturbance information, calculating the change probability of wetland change and the change angle of the change disturbance information, and carrying out change judgment according to the change probability and the change angle to obtain a judgment result;
and performing difference comparison on the fitted wetland band data and the denoised wetland band data to obtain change disturbance information, calculating the change probability of wetland change and the change angle of the change disturbance information, and performing change judgment according to the change probability and the change angle to obtain a judgment result, wherein the judgment result comprises:
performing difference comparison on the fitted wetland band data and the denoised wetland band data to obtain the variation disturbance information of the denoised wetland band data;
calculating the probability of wetland change based on the difference value between the fitted wetland band data and the denoised wetland band data to obtain the change probability value, and comparing the change probability value with a dynamic threshold value;
calculating the change angle of the change disturbance information to obtain a change angle value, and comparing the change angle value with an angle threshold value;
carrying out change judgment according to the change probability and the change angle to obtain a changed judgment result and an unchanged judgment result;
and performing change judgment according to the change probability and the change angle to obtain a judgment result, wherein the step of obtaining the judgment result comprises the following steps of:
if the change probability value is larger than the dynamic threshold value and the change angle value is larger than the angle threshold value, the change is judged, and a change judgment result is obtained;
if the change probability value is smaller than the dynamic threshold value and/or the change angle value is smaller than the angle threshold value, judging that the change is unchanged, and obtaining an unchanged judging result;
the step of selecting the change disturbance information according to the judging result to obtain wetland time sequence change information comprises the following steps:
if the unchanged judging result is obtained, selecting not to screen the changed disturbance information, and taking the changed disturbance information as the wetland time sequence change information;
if the judging result of the change is obtained, judging the part of the change disturbance information which is not clear through the slope trend direction, determining final change disturbance information, removing vegetation green degree increasing information caused by vegetation green and artificial factors in the change disturbance information, obtaining wetland time sequence change information, and taking the wetland time sequence change information as a change detection result;
and the disturbance information extraction module is used for selecting the change disturbance information according to the judging result to obtain wetland time sequence change information.
6. A computer-readable storage medium storing one or more programs executable by one or more processors to implement the steps in the method for detecting a time-series change of a wetland of any one of claims 1 to 4.
7. A terminal device, comprising: a processor, a memory, and a communication bus; the memory has stored thereon a computer readable program executable by the processor;
the communication bus realizes connection communication between the processor and the memory;
the processor, when executing the computer readable program, implements the steps in the method for detecting time-series changes of a wetland according to any one of claims 1 to 4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211152897.0A CN115795255B (en) | 2022-09-21 | 2022-09-21 | Method, device, medium and terminal for detecting time sequence change of wetland |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211152897.0A CN115795255B (en) | 2022-09-21 | 2022-09-21 | Method, device, medium and terminal for detecting time sequence change of wetland |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115795255A CN115795255A (en) | 2023-03-14 |
| CN115795255B true CN115795255B (en) | 2024-03-26 |
Family
ID=85432087
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211152897.0A Active CN115795255B (en) | 2022-09-21 | 2022-09-21 | Method, device, medium and terminal for detecting time sequence change of wetland |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115795255B (en) |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107895169A (en) * | 2017-10-25 | 2018-04-10 | 南京邮电大学 | A kind of method based on ENVISAT ASAR dual polarizations data extraction wetland information |
| CN108229426A (en) * | 2018-01-29 | 2018-06-29 | 南京信息工程大学 | A kind of Remote Sensing Imagery Change vector based on difference description changes detection method |
| CN108564002A (en) * | 2018-03-22 | 2018-09-21 | 中国科学院遥感与数字地球研究所 | A kind of remote sensing image time series variation detection method and system |
| CN110751727A (en) * | 2019-10-30 | 2020-02-04 | 南京林业大学 | Synthetic image construction method based on Landsat long-time sequence |
| CN112053359A (en) * | 2020-09-30 | 2020-12-08 | 中国科学院空天信息创新研究院 | Remote sensing image change detection method and device, electronic equipment and storage medium |
| CN112990556A (en) * | 2021-02-24 | 2021-06-18 | 江苏大学 | User power consumption prediction method based on Prophet-LSTM model |
| CN113160359A (en) * | 2021-05-21 | 2021-07-23 | 宁波大学 | Coastal wetland remote sensing mapping method based on harmonic fitting parameters |
| CN114120137A (en) * | 2021-10-19 | 2022-03-01 | 桂林理工大学 | Wetland element space-time evolution monitoring method based on time sequence main remote sensing image |
| CN114155429A (en) * | 2021-10-09 | 2022-03-08 | 信阳学院 | Reservoir earth surface temperature prediction method based on space-time bidirectional attention mechanism |
| CN114359735A (en) * | 2021-12-31 | 2022-04-15 | 中国科学院空天信息创新研究院 | Hyperspectral remote sensing image change detection method |
| CN115082784A (en) * | 2022-05-31 | 2022-09-20 | 同济大学 | Multi-temporal remote sensing image unsupervised change detection pseudo sample automatic generation method |
-
2022
- 2022-09-21 CN CN202211152897.0A patent/CN115795255B/en active Active
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107895169A (en) * | 2017-10-25 | 2018-04-10 | 南京邮电大学 | A kind of method based on ENVISAT ASAR dual polarizations data extraction wetland information |
| CN108229426A (en) * | 2018-01-29 | 2018-06-29 | 南京信息工程大学 | A kind of Remote Sensing Imagery Change vector based on difference description changes detection method |
| CN108564002A (en) * | 2018-03-22 | 2018-09-21 | 中国科学院遥感与数字地球研究所 | A kind of remote sensing image time series variation detection method and system |
| CN110751727A (en) * | 2019-10-30 | 2020-02-04 | 南京林业大学 | Synthetic image construction method based on Landsat long-time sequence |
| CN112053359A (en) * | 2020-09-30 | 2020-12-08 | 中国科学院空天信息创新研究院 | Remote sensing image change detection method and device, electronic equipment and storage medium |
| CN112990556A (en) * | 2021-02-24 | 2021-06-18 | 江苏大学 | User power consumption prediction method based on Prophet-LSTM model |
| CN113160359A (en) * | 2021-05-21 | 2021-07-23 | 宁波大学 | Coastal wetland remote sensing mapping method based on harmonic fitting parameters |
| CN114155429A (en) * | 2021-10-09 | 2022-03-08 | 信阳学院 | Reservoir earth surface temperature prediction method based on space-time bidirectional attention mechanism |
| CN114120137A (en) * | 2021-10-19 | 2022-03-01 | 桂林理工大学 | Wetland element space-time evolution monitoring method based on time sequence main remote sensing image |
| CN114359735A (en) * | 2021-12-31 | 2022-04-15 | 中国科学院空天信息创新研究院 | Hyperspectral remote sensing image change detection method |
| CN115082784A (en) * | 2022-05-31 | 2022-09-20 | 同济大学 | Multi-temporal remote sensing image unsupervised change detection pseudo sample automatic generation method |
Non-Patent Citations (3)
| Title |
|---|
| Diversity of Algorithm and Spectral Band Inputs Improves Landsat Monitoring of Forest Disturbance;Warren B. Cohen等;remote sensing;第12卷(第10期);第1-15页 * |
| Landsat 时序变化检测综述;汤冬梅等;地球信息科学学报;第19卷(第8期);第1069-1079页 * |
| 一种改进的基于RNN的遥感影像变化检测方法研究;曹州;中国优秀硕士学位论文全文数据库，基础科技辑(第6期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115795255A (en) | 2023-03-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Seebens et al. | Predicting the spread of marine species introduced by global shipping | |
| Schepaschenko et al. | Development of a global hybrid forest mask through the synergy of remote sensing, crowdsourcing and FAO statistics | |
| Béjaoui et al. | Random forest model and TRIX used in combination to assess and diagnose the trophic status of Bizerte Lagoon, southern Mediterranean | |
| Topp et al. | Multi-decadal improvement in US Lake water clarity | |
| Martínez Cano et al. | Tropical tree height and crown allometries for the Barro Colorado Nature Monument, Panama: a comparison of alternative hierarchical models incorporating interspecific variation in relation to life history traits | |
| Dye et al. | Examining the utility of random forest and AISA Eagle hyperspectral image data to predict Pinus patula age in KwaZulu-Natal, South Africa | |
| Júnior et al. | Machine learning: Modeling increment in diameter of individual trees on Atlantic Forest fragments | |
| De Marco et al. | Vulnerability of Cerrado threatened mammals: an integrative landscape and climate modeling approach | |
| García-Martínez et al. | The surrounding landscape influences the diversity of leaf-litter ants in riparian cloud forest remnants | |
| Mellor et al. | Exploring diversity in ensemble classification: Applications in large area land cover mapping | |
| Lundström et al. | Lichen species richness on retained aspens increases with time since clear-cutting | |
| Yamakita et al. | Identification of important marine areas around the Japanese Archipelago: Establishment of a protocol for evaluating a broad area using ecologically and biologically significant areas selection criteria | |
| CN114740550B (en) | Intelligent recognition early warning method and system for continuous storm events | |
| Bulut | Machine learning prediction of above-ground biomass in pure Calabrian pine (Pinus brutia Ten.) stands of the Mediterranean region, Türkiye | |
| Guo et al. | Half of the world’s tree biodiversity is unprotected and is increasingly threatened by human activities | |
| Mazor et al. | Eutrophication thresholds associated with protection of biological integrity in California wadeable streams | |
| CN115795255B (en) | Method, device, medium and terminal for detecting time sequence change of wetland | |
| Cwynar et al. | Depth distribution of chironomids and an evaluation of site-specific and regional lake-depth inference models: a good model gone bad? | |
| Hsu et al. | Evaluation of the influence of spatial treatments on catch-per-unit-effort standardization: A fishery application and simulation study of Pacific saury in the Northwestern Pacific Ocean | |
| Martins et al. | Successional convergence in experimentally disturbed intertidal communities | |
| Hallstan et al. | Comparison of classification-then-modelling and species-by-species modelling for predicting lake phytoplankton assemblages | |
| Van Nguyen et al. | Multi-task convolution neural network for season-insensitive chlorophyll-A estimation in inland water | |
| Brasil et al. | Contributions of Forest Regeneration After Intense Fragmentation in the Amazon through Morphological Spatial Pattern Analysis | |
| Zeng et al. | Contrasting diversity patterns of breeding Anatidae in the Northern and Southern Hemispheres | |
| CN113742438B (en) | Method and device for determining landslide susceptibility distribution map and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |